Unsupervised Detection of Lesions in Brain MRI using constrained adversarial auto-encoders

Abstract

Lesion detection in brain Magnetic Resonance Images (MRI) remains a challenging task. State-of-the-art approaches are mostly based on supervised learning making use of large annotated datasets. Human beings, on the other hand, even non-experts, can detect most abnormal lesions after seeing a handful of healthy brain images. Replicating this capability of using prior information on the appearance of healthy brain structure to detect lesions can help computers achieve human level abnormality detection, specifically reducing the need for numerous labeled examples and bettering generalization of previously unseen lesions. To this end, we study detection of lesion regions in an unsupervised manner by learning data distribution of brain MRI of healthy subjects using auto-encoder based methods. We hypothesize that one of the main limitations of the current models is the lack of consistency in latent representation. We propose a simple yet effective constraint that helps mapping of an image bearing lesion close to its corresponding healthy image in the latent space. We use the Human Connectome Project dataset to learn distribution of healthy-appearing brain MRI and report improved detection, in terms of AUC, of the lesions in the BRATS challenge dataset.

9 pages, 5 figures, accepted at MIDL 2018

References in corpus (3)

• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Improved Training of Wasserstein GANs
• Deep Unsupervised Clustering with Gaussian Mixture Variational Autoencoders

Cited by in corpus (17)

• Generative Adversarial Network in Medical Imaging: A Review
• Deep Learning Approaches for Data Augmentation in Medical Imaging: A Review
• Anomaly Detection in Medical Imaging with Deep Perceptual Autoencoders
• EEGFuseNet: Hybrid Unsupervised Deep Feature Characterization and Fusion for High-Dimensional EEG with An Application to Emotion Recognition
• Dual-distribution discrepancy with self-supervised refinement for anomaly detection in medical images
• Artificial Intelligence in Glioma Imaging: Challenges and Advances
• Pseudo-healthy synthesis with pathology disentanglement and adversarial learning
• Detecting Outliers with Foreign Patch Interpolation
• Learning joint segmentation of tissues and brain lesions from task-specific hetero-modal domain-shifted datasets
• Anomaly detection and automatic labeling for solar cell quality inspection based on Generative Adversarial Network
• Constrained domain adaptation for Image segmentation
• Neural Transformation Learning for Deep Anomaly Detection Beyond Images
• Evaluation of pseudo-healthy image reconstruction for anomaly detection with deep generative models: Application to brain FDG PET
• Improving unsupervised anomaly localization by applying multi-scale memories to autoencoders
• Beyond Individual Input for Deep Anomaly Detection on Tabular Data
• Progressive GANomaly: Anomaly detection with progressively growing GANs
• Quantum Patch-Based Autoencoder for Anomaly Segmentation